Özet:

COVID-19 pandemic affecting the whole world, reveals the importance of the studies that trying to detect the outbreaks in early stage. If any outbreak can be detected in an early stage, the number of infected people can be reduced, the necessary treatment can be found and treatment expenses can be also reduced. The most important data processing approaches enabling to detect outbreaks in an early stage are machine learning approaches, which use mathematical models and statistical background. With machine learning techniques, medical data can be analyzed and processed to make predictions of illnesses. Because, previously collected patient datasets help to perform these predictions. Beside illnesses, outbreaks can be also predicted by using these collected datasets. Machine learning techniques enable us to process labelled and unlabelled datasets with the help of supervised and unsupervised approaches, respectively. Although there are many supervised learning approaches like Naïve Bayes (NB), Support Vector Machine (SVM), Decision Trees (DT), Random Forest (RF) and Artificial Neural Networks (ANN) to predict the emergence of the outbreaks that appeared before, it is also possible to detect any outbreak which are unprecedented before by using unsupervised learning approaches like principal component and cluster analysis. In this study, it is aimed to present a detailed analysis of machine learning approaches in outbreak detecting area to give a lead to the researchers who want to work in this area.

Anahtar Kelimeler: